Contruction of the Cultural Homogeneity in VC Dataset

Jump to navigation Jump to search

This page details the (currently 1st draft version) construction of the Cultural Homogenteity in Venture Capital Data. This data is posted for reference (and debate) by project members. The choices made in constraints, variables, processing, and so forth are often arbitrary and will no doubt be changed as the project progresses.

Constraints on Retrieved Data

The following constraints were placed in the retrieval of data from VentureXpert (via SDC Platinum):

  • Portfolio Company Date of First Investment: 2003-2007 inclusive
  • Portfolio Company Nation: US
  • Venture Capital Fund Nation: US
  • Portfolio Company Standard US Venture Disbursement: Yes (Note: Correlates almost perfectly with PWC Moneytree VC Deals - excludes private equity, angel investment, and other non-VC, as recognised by Thomson)
  • Venture Capital Fund PWC Moneytree Deals: Yes (see above)

Retrieved Data by VentureXpert Perspective

The data was processed by the perl script to produce 3rd normal form relational database tables. In the case of executive titles this was problematic and required rekeying in the database.

Portfolio Companies

  • Company (Primary Key: CoExex-1)
    • CoExec-1, DateCompanyReceivedFirstInvestment, CompanyFoundingDate, CompanyZipCode, CompanyCity, CompanyAreaCode, CompanyCounty, CompanyName, CompanyNation, CompanyNationCode, CompanyState, CompanyStateCode, CompanyStreetAddressLine1, CompanyStreetAddressLine2, CompanyIndustryClass, CompanyIndustryMajorGroup, CompanyIndustryMinorGroup, CompanyIndustrySubGroup1, CompanyIndustrySubGroup2, CompanyIndustrySubGroup3, StandardUSVentureDisbursement
  • CoExec (Primary Key: Co-Exec-3)
    • CoExec-1, CoExec-3, ExecutiveisNonManagingBoardMember, ExecutiveisPrimaryContact, ExecutivesCity, ExecutivesFirstName, ExecutivesLastName, ExecutivesNamePrefix, ExecutivesEMailAddress, ExecutivesPhoneNumber, ExecutivesPreviousPosition
  • CoExecTitle (Primary Key: Co-Exec-2)
    • CoExec-1, CoExec-2, ExecutivesJobTitle


  • Fund (Primary Key: FundExec-1)
    • FundExec-1, FirmFundList1stCloseDateofeachFund, FirmFoundingDate, FundInitialClosing, FundAreaCode, FundCity, FundCounty, FundStageFocus, FundInvestmentType, FundMSA, FundMSACode, FundName, FundNation, FundNationCode, FundRaisingStatus, FundSequenceNo, FundSequenceType, FundSizeMil, FundState, FundStateCode, FundTargetSize000, FundTypeLongDescription, FundTypeShortDescription, FundYear, FundZipCode, FirmReportedCapitalunderMgmtMil, FirmNation, FirmName, FirmGeographyPreference, FirmIndustryPreference, FirmPreferredInvestmntRoleCode, FirmInvestmentStagePreference, FirmPreferredMaxInvestmentMil, FirmPreferredMinInvestmentMil, FirmState, FirmZipCode, FirmStateCode, PWCMoneytreeDealsYN, FirmFundListNameofeachFund
  • FundExec (Primary Key: FundExec-2)
    • FundExec-1, FundExec-3, ExecutiveisNonManagingBoardMember, ExecutiveisPrimaryContact, ExecutivesCity, ExecutivesFirstName, ExecutivesLastName, ExecutivesNamePrefix, ExecutivesPreviousPosition
  • FundExecTitle (Primary Key: FundExec-2)
    • FundExec-1, FundExec-2, ExecutivesJobTitle


  • Round (Primary Key: Round-2)
    • Round-1, Round-2, RoundDates, CompanyStageLevel1ateachRoundDate, RoundAmtDisclosed000, RoundAmtEstimated000, RoundNumbers, NumberofInvestorseaRound
  • RoundCompany (Primary Key: Round-1)
    • Round-1, CompanyName, CompanyState, CompanyStateCode, TotalKnownAmtInvestedinCompany000
  • RoundInvestor (Primary Key: Round-3)
    • Date, DisclosedAmtk, Investor, Round-1, Round-3

Lookup Tables

Lookup tables for Stage, State and VentureXpert Minor Industry Code are provided on the VentureXpert page. The numeric values were recorded in the database.

Processing the Data

The following problems were addressed and steps undertaken:

  • Executives (both Company and Fund executives) have multiple job titles - the first listed title was taken (requires rekeying)
  • Executive Name Prefixes were parsed for Dr and Ms (recorded as binary codes)
  • Anglicized First Names and genders were (incompletely) identified using the U.S. Social Security Administration Baby Names list
  • Company Executives with the following SQL constraint were marked as founders:
    • Like "*Founder*" Or Like "*Chief Exec*" Or Like "*CEO*" Or (Like "President*" Or "Co-President") Or Like "*Chairman*"
  • Company Executives marked as "N" for Non-Managing Board Members were coded as Managing
  • Fund Executives meeting the following SQL constraint were marked as "Partners" (broadly defined):
    • (Like "*Partner*" And Not Like "Admin*" And Not Like "Advisor*" And Not Like "Limited*") Or "Principal" Or "Director" Or "Managing Director" Or "CEO" Or Like "*Founder*" Or Like "*Investment*" Or Like "*Executive Officer*"
  • Fund Executives were marked as ServedOnBoard if they appeared on the board of any portfolio company in the dataset at any time
  • The Company and fund zip codes were standardized to 5 digits
  • Companies must not have null or "Undisclosed" names
  • Funds must not be named "Undisclosed Fund" and must have participated in a first round of investment in the period of interest (i.e. within the constraint on date of first investment on companies)

Output Files

Based on the above processing, the following comma-seperated, text-quoted, text files were created (Note requires ftp login):